MYC Amplification As a Prognostic Biomarker in Osteosarcoma

MYC Amplification As a Prognostic Biomarker in Osteosarcoma
Dr. Marinoff graduated from Harvard Medical School and completed her pediatric residency at Boston Combined Residency Program, during which time she worked with Dr. Katie Janeway on elucidating the genomic landscape of and novel genomic biomarkers in osteosarcoma. She is currently completing her fellowship training in Pediatric Hematology/Oncology at UCSF, where she is focused on developing novel genome-informed therapeutics for osteosarcoma under the mentorship of Dr. Alejandro Sweet-Cordero. She plans to develop an active clinical and translational research program focused on conducting early phase precision medicine-oriented trials for patients with advanced sarcomas. She is grateful to have the best job in the world.

For all past and future episodes of OsteoBites, click here !

Christina Ip-Toma: Welcome toOsteoBites. My name is Christina Ip-Toma and I am mom to Osteo Angel Dylan andI'm the director of scientific programs for MIB agents. Today on OsteoBites,we're talking with Dr. Amanda Marinoff from UCSF, who will be discussing her research on MYC amplification as a first genomic prognostic biomarker and osteosarcoma, and how it may be used for risk stratification in future clinical trials to inform conversations with patients and families, and how it may serve as a potential therapeutic target in the future.

Thank you so much, Dr. Marinoff, for joining us on osteophytes today, Dr.Marinoff graduated from Harvard Medical School and completed her pediatric residency at Boston combined residency program, during which time she worked with Dr. Katie Janeway on elucidating the genomic landscape of and novel genomic biomarkers in osteosarcoma. She's currently completing her fellowship training in pediatric hematology-oncology at UCSF, where she's focused on developing novel genome-informed therapeutics for osteosarcoma under the mentorship of Dr. Alejandro Sweet-Cordero. She plans to develop an active clinical and translational research program focused on conducting early phase precision medicine-oriented trials for patients with advanced sarcomas.

Before we get started with Dr. Marinoff's presentation, I just have a few events that I wanted to share with you all. First, if you're in Boston, right after this podcast you need to get on the T, take the green line to Arlington and head straight to Davio's Back Bay for free drinks and snacks, and end the day with some good vibes hanging out with it MIB agents crew and the Boston osteosarcoma community. We have another event in Boston tomorrow. It's not too late to sign up for the Healing Hearts retreat for bereaved osteosarcoma patients and parents on Friday in Boston. Email Casey to get more info and we'll put her email address in the chat. Last announcement is Factor registration is now open. So, we'll put the registration link in the chat also.It's going to be June 23rd to 25th in San Diego. And you know what they say about Vegas, you know what happens in Vegas, stays in Vegas? So, Factor is the opposite of that. What happens at Factor has ripple effects way beyond and you're going to get a dose of inspiration and hope that will last long after San Diego. So, please come, you will be very happy you did.

Before we start, I just want to thank our sponsor for this episode BTG Specialty Pharmaceuticals. BTG provides rescue medicines typically used in emergency rooms and intensive care units to treat patients for whom there are limited treatment options. They are dedicated to little varying quality medicines that make a real difference to patients and their families through the development, manufacture and commercialization of pharmaceutical products.Their current portfolio of antidotes counteracts certain snake venoms and the toxicity associated with some heart and cancer medications. Their drug Voraxazis for high dose methotrexate toxicity. So, with that, Dr. Marinoff, if you want to go ahead, you can share your presentation and we'll get started.

Dr. Amanda Marinoff: Thank you so much for that introduction, Christina and for having me today on the webinar. It is really such an honor to be here. As Christina shared, I'm a pediatric oncology fellow and taking care of pediatric patients with cancers and thinking about how to improve their care is really the best job in the world. So, I'm excited to talk to everyone today about how we can understand MYC as a way to make it better for children with osteosarcoma.

I'll share first that I developed my interest in osteosarcoma during my time in medical school, while I was on my surgery rotation, which quite honestly was not my favorite but ended up being a very formative experience because I met avery remarkable young man during his first open chest surgery to remove metastatic osteosarcoma nodules from his lung. When he woke up and in the ensuing months, he was later introduced to me as to his nurses and doctors ashey, this is Amanda. She touched my heart. No, literally, she was inside my chest. I will say Miles is his name. From a metaphorical standpoint, certainly touched mine and both his life and his ultimate passing inspired my interest in pediatric oncology in general, and my interest in studying osteosarcoma specifically. So, with that, I do have a couple of disclosures. First, I want to make it better for children with osteosarcoma and their families. Secondly,I believe we can and we will.

So, just a brief bit of background, this group likely doesn't need a whole lot,just to sort of send our ROI. Over the past four decades, we've had limited progress in terms of treatment strategies, outcomes and our risk stratification strategies. This is a diagram, a busy one that shows the survival curves based on the last four decades of clinical trials. The punchline here is, we haven'tmade much progress, but it hasn't been for lack of trying. We've seen each ofthese lines is a different trial. The green ones include four differentchemotherapy agents, the blue ones include three different chemotherapy agents,and the red ones include two. We see the green and blue do a little bit betterthan the red but not much better than each other. And that more chemotherapyisn't better.

So, our standard of care has been, in North America at least, three chemotherapy agents map methotrexate, adriamycin, and platinum is what itstands for. Patients get chemotherapy upfront, then they get surgery to remove their primary tumor, and then they get more map therapy. So, more agents from astandard chemotherapy point of view has not resulted in better outcomes. So, we need to think outside the box, we need to better understand the biology andcome up with different strategies.

So, that affords us a number of clinical challenges and opportunities, thefirst of which is our very poor outcomes specifically for patients withadvanced osteosarcoma. And those are those with relapsed or metastatic disease.We have blunt and imperfect tools for risk stratification, we use the presenceof metastasis to define high risk disease if patients weren't able to have acomplete resection of their primary tumor. Then we look at histologic responseto chemotherapy, meaning when patients have their tumor removed surgically, thepathologist looks at it under the microscope and says what percentage of thecells are dead or necrotic, and as a proxy for its response to chemotherapy.But that's not a perfect tool, either. There are some patients who have a poorresponse histologically who do well, and and vice versa.

So, we're not using any molecular markers in our re stratification schema. Andeven when we can identify high risk patients, we don't actually have therapiesavailable yet to to intensify or target their specific disease and there's notargeted therapies that were yet integrating into our regular clinical care.And on top of that, osteosarcoma has very complex genomics and is an extremelyheterogeneous disease. But with these challenges, Houston I think we have anumber of opportunities to just briefly review the chaotic genome that we havenow elucidated through a number of landmark sequencing studies over the lastseveral years.

Osteosarcoma is characterized by widespread copy number alterations, there aresome pediatric cancers where we have sort of a single Fusion Driver event thatwe know to be present in almost all patients with with that disease subset. Butthat's not at all the case for osteosarcoma. This first top plot is what wecall a copy number plot. The numbers on the x-axis are just chromosome numbers.In a normal genome, we would not see any of those blue valleys or those yellowand red peaks, it would be quiet, but we see here that there are a number ofcopy number losses shown in blue, and then GAIN shown in yellow and red allthroughout the genome. Then these circles at the bottom are pretty but areevidence of more genomic chaos. We call these circles plots. They show an eventthat we call chromothripsis, which is basically taking the genome and thenshattering it and causing a number of widespread changes from a single event.We can see first and foremost that there are different colors within each plotand then compared to one another, and osteosarcoma has been scribed as sort oflike a snowflake, one physician-scientist said to me, if you have a genome thatlooks the same from one patient to another, you've probably contaminated yoursamples. That's how heterogeneous this disease is. We see here that the purplelines indicate translocations orange lines are inversions. green lines areduplications, and blue lines are deletions.

So, our question and our task is, can we find some signal amid the noise to define clinically relevant subgroups that will allow us to predict how patients will do, how their disease will behave, who will respond to standard therapy and who needs different therapy up front and not just that relapse that really fits with their disease biology? So, with that, I'll share some data from a project I worked on under the mentorship of Dr. Katie Janeway at Dana Farber which we looked at a high risk pediatric cohort of patients with osteosarcoma that were from two studies. One was the profile study which sequences all patients that are seen at Dana Farber with cancer diagnoses. Then a second is from the GAIN cohort. It stands for Genome Assessment Informs Novel therapy.It's also called the iCat2 study and that enrolled patients from 12 different institutions who had high risk relapsed or metastatic extracranial solid tumors.

So, we knew that this was a relatively high risk cohort, just by virtue oftheir enrollment on the GAIN study. That was on purpose, because this is reallythe group that we think needs a better understanding and for which we needbetter therapies. So, we first looked at their survival curves overall, andtook a look at the different clinical prognostic markers to see if there was any signal for clinical factors that could differentiate subgroups.Unsurprisingly, we see a difference in terms of patients who had localizeddisease, shown in yellow versus those who had metastatic disease, as shown ingray, and the overall cohort is shown in blue. We see that the three yearoverall survival for those with localized disease was 77%, whereas those with metastatic disease was 45%. We see a similar pattern in terms of event-freesurvival. Of all the clinical prognostic factors that we looked at the presence of metastasis was the only one that was statistically significant, including thepercent of necrosis or the response to chemotherapy, when we look at those tumors under the microscope did not turn out to be statistically significant.

So, then we said, can we actually take a look at their genomic landscape? For a couple of reasons. One, this is a group of relatively advanced high risk,osteosarcoma, do their alterations look any different than kind of osteosarcoma[inaudible]? Secondly, can we identify any biomarkers that can be used for risk stratification, and or a potential therapeutic target. This is a busyonco-print, which shows the most commonly altered genes in our cohorts, shown on the left over here, along with their clinical characteristics up top. Each of these bars is a different patient with a different alteration. This sequencing was done based on targeted panel-based sequencing called onco-panelusing a next-generation sequencing capture-based method. That's a clinical tool that pretty much every patient who gets treated on or off trial gets sequence with. We saw here that MYC was amplified and 12% of patients with osteosarcoma.That's somewhat similar to what's reflected in the literature, which has a somewhat broad range from somewhere in the 10 to 40% range of patients with MYC amplifications, depending on the sequencing method and the cohort that's used.

Christina: I just grabbed a couple questions here. So, for that, do you knowbecause I mean, MYC is is kind of a common mutation in other more commoncancers, and do we have a sense of kind of how that 12% compares to othercancers?

Dr. Marinoff: Yeah, that's a great question. MYC is not a new gene that we know of. It's not unique to osteosarcoma. In fact, MYC related pathways are thought to be dysregulated and about 70% of cancers. A recent pan-cancer analysis using the TCGA cohort, looking at 9000 samples found that it was amplified specifically and 28% of cancers overall, but probably breaks down differently in terms of different cancer subsets. Then within cancers, I think probably different subsets with it within there. It depends on how you define MYC alteration and what methodology you use to sequence it. Overall, it's a lot of cancers across the board.

Christina: Right, and just this list is kind of prompting, you might be gettingto this later but I'm just curious. When you see the MYC amplification, is itcommonly paired with or showing up with other mutations? Like does it tend toshow up with other mutations or is it just kind of its own thing?

Dr. Marinoff: That's a great question. And I think it's one that we're stillexploring to look at which, which genes are co-amplified, what that meansprognostically for our patients, and which ones with which it's mutuallyexclusive. We saw a signal here and that's been shown in a couple of othersmall studies. If you look at MYC and CCNE1, they seem to be tracking togetherand there was a recent paper by De Noon, a group that showed that thosepatients seem to have a particularly poor prognosis. But I think that remainsto be validated and other interactions with MYC is a really important question too, for us to to explore.

So, of all of the genomic markers that we looked at, MYC amplification was the only one that was statistically associated with survival in our cohort. That's shown here in our Kaplan Meier curves. The top curves are overall survival in A and event-free survival. B, for our primary analytic cohort, those are the 92 patients that I just described. We see in blue those with-- sorry, in red,those would MYC amplification defined at a copy number, a threshold of greater than seven, and have a significantly worse overall survival and attend toward worse event-free survival than those without MYC amplification are those with less than or equal to seven copies of MYC. I'll talk on our next slide about kindof how we chose that threshold, why it's good, why it's not great, and where we go from there.

We also looked at a validation cohort from 86 patients from Memorial Sloan Kettering because we were wondering, hey, this is a high risk group of patients, is MYC really just like we have more patients with metastatic disease. Is that what's going on here? Is that just a proxy? What if we looked at a more homogenous group of patients that are more kind of standard risk in localized patients only? So, these are 86 patients with localized disease.Indeed, these patients, those with among those with MYC amplification shown in red, and in here MYC amplification is defined differently as a two-fold change or greater, did significantly worse than those without MYC amplification from both an overall survival standpoint, and an event-free survival standpoint.

Now, clearly, these are not exactly comparing apples to apples, we're using different sequencing methodologies, then Memorial Sloan Kettering use a different targeted sequencing base called MSK-Impact. We have different definitions of MYC amplification so it's really hard to to completely synthesize all this data. However, I do think that the fact that we had two distinct and independent cohorts, using different bioinformatics pipelines,using different sequencing tools, different definitions, what we're still kind of seeing at the crux of it is MYC amplification does seem to have an association with poor outcomes in both of these cohorts, and I think really, it calls to the robustness of that observation, and, and a call to continue to explore and validate that.

Which cut off made the cut? Well, I told you, we chose copies greater than 7,and we did an analysis, looking at association, each of these copy number thresholds.You can see here, actually, a copy number is greater than 4, we do see a statistical significance in terms of association with survival. P here is less than 0.05, but we see that this association doesn't hold up at copies greater than 5 and 6 at those thresholds, potentially suggesting that at these lower copy number amplifications, that there's a lower signal to noise ratio. Butthen at 7, and then greater than 8, 9, and 10, we see consistently a significant association with poor outcomes with MYC amplification. So, again,we chose our copy number threshold at 7. This is start here on blue, it's not yet a gold start, because again, it needs to be validated in additional cohorts, ideally using the same sequencing and bioinformatics pipelines and ina more homogeneously analyzed group.

One question that sometimes comes up is, is more MYC worse? I think we don'tyet know that. I think there's a lot of layers of complexity, but I'll justshare a few clinical anecdotes that I think are at least intriguing and merit further looking into this question about how the number of copies of MYC mayrelate to the the disease's behavior and the clinical outcomes for our patients. So, we saw three patients with extremely high MYC amplification. They had particularly aggressive courses. In this case of patient six, we see an Aand the copy number plot with absolute copy number here on the y axis and wesee actually almost 80 copies of MYC we were just talking about in the four,maybe four range, five, six, seven, this is 80 copies of MYC. This patient presented at a younger than average age, usually we see preteens andadolescents is patient presented at the age of 7. Then B we see kind of a snapshot of their PET scan and all of these black spots are our areas of metastasis.These are abnormal areas that light up on the PET scan and show diffusely bony metastatic disease in addition to lung metastases in this very high MYCamplified osteosarcoma.

This MYC amplification was actually detected in circulating tumor DNA incirculating tumor cells, which is not yet something that we have fullyintegrated into clinical practice. But we had there, the suggestion here thatMYC amplification could be detected. This is something that could potentiallybe validated be used to track our response to therapy, and then minimalevidence of relapse before frank relapse. Again, this patient had remarkablymetastatic disease presentation and then progressed after four cycles of ourstandard map therapy, which is quite unusual to progress while on therapy, andthen ultimately died 10 months after presentation.

In a second case, in case 24, we see MYC is amplified to a little bit over 30copies. This patient presented with initially localized disease at the age of16. Based on our current risk stratification schema, we would say that patientswith localized disease tend to have relatively favorable outcomes, with overallsurvival in the 60 to 70% range. However, this patient had a recurrence sixmonths after completing their map chemotherapy, and then died within threeyears of presentation. Then similarly, this case 32 had a MYC number levelaround 30 and this patient also presented very young at the age of 6. Thispatient had localized disease, and so was thought to potentially have relativelyfavorable outcome, but had a poor response to chemotherapy based on looking attheir tumor cells under the microscope, and then died of disease a year afterthey presented. So, again, all of these are very much anecdotal, but I thinkare outliers, and that warrant, that there are sort of food for thought andfruit for further investigation in terms of the relationship between the numberof copies of MYC and the disease presentation.

Christina: Dr. Marinoff, just so I'm understanding so, is the x-axis therethat's kind of longitudinal over time from like diagnosis over the course oftheir treatment. Is that correct?

Dr. Marinoff: Copy number plot? Mm-hmm. Great, great question, and sorry fornot clarifying. That is just across the chromosomes and where were MYC glands.So, the x-axis doesn't really have specific relevance. It's really that y-axisshowing the peak in terms of the absolute copy numbers. I'll talk a little bitlater about some emerging data indicating that MYC copy number at the genomiclevel really isn't the full story and that we need to be thinking about otherlevels of regulation, including the way that it's expressed and the way thatit's epigenetically regulated.

Christina: Because I'm curious if you have data showing like longitudinal datawhere you can kind of see at diagnosis, but then with progression andmetastatic disease, if you're seeing like, is it common to see kind of MYCappear at diagnosis, and then perhaps like the copy number changes, likeincrease over time, or like, can you not have any MYC amplification adiagnosis, but then two years down the line...

Dr. Marinoff: I have a slide toward the end, but I'll share it now. Thanks forasking that. I think the question is right on the money. We really need tounderstand the genomic evolution of osteosarcoma, and the relationship betweenprimary tumor and metastatic tumor. That is really through collectinglongitudinal serial samples over time and matching the initial primary tumor totheir later recurrent and-or metastatic samples. Thus far in the literature, wedon't have a lot of that data. But I know there's a lot of interest in it. Dr.Jamie and I did an exploratory analysis. In our dataset of 92 patients, we had13 patients who had multiple samples, they weren't all perfectly matched ofprimaries to their metastatic tumors. We at least got to start to dabble insome patterns. We see here MYC up top again, each of these vertical columns arean individual patient.

We can see in some cases, MYC was present, surprisingly in the primary tumorsample, but seem to have disappeared in the metastatic sample and other casesand darkness is kind of correlates with a number of copies, whereas lighter isfewer numbers of copies, we had fewer copies that were detected. In other cases,in the metastatic sample we had more copies detected, so it's a little bit allover the map and I think remains to be explored. I think part of this could betechnical issues, it could just be the way that we're looking at it anddetecting mech, it could be related to tumor purity, but some of it couldpotentially be a real biological phenomena. There is some evidence in at leastin the breast cancer world that, that breast cancer cells downregulate MYC, atleast transiently to mediate metastasis. So, I think it's tempting to say thatmore MYC equals more metastasis and draw that line directly from A and B, but Ithink it's probably not that straightforward. Doing further studies and alarger cohorts and really matched samples is going to really help to disentanglethat, to understand that genomic evolution and other in other genomicalterations.

I'll just share here and so my interest in MYC developed from that study that Idid with Katie, but there's been some, a lot of emerging evidence recently overthe past few years and a few different cohorts indicating the same associationof, of MYC amplification with poor outcomes. This was a study from Chen et althat looked at 30 different tumors and looked at whether MYC expression wasmore prevalent in patients who develop metastatic disease. We see here this redbox and whisker plot tends to have higher levels of MYC levels of MYC and a logfashion is on the y axis than those who did not develop metastatic disease.Another way of looking at that is through the survival curve, where patientswith high MYC amplification shown in blue have a lower likelihood of metastaticfree survival.

Similarly, a study by Chen et al which looked at 70 different tumors from atumor microarray and then validated it with a publicly available data setshowed that those with MYC high expression here in yellow, had poor overall andprogression-free survival relative to low MYC shown in pink and then De Noon etal looked at three additional publicly available datasets and found that MYCamplification actually was the one signal that differentiated pediatric tumorsor was enriched in pediatric osteosarcomas, relative to their adultcounterparts. And again, at least in the localized setting, they didn't haveenough numbers exactly in the metastatic MYC amplified cohort, but MYCamplified localized osteosarcoma does just as bad as the metastatic group withMYC wild type and much worse than those who have no MYC amplification andlocalized disease.

So, with all of this, I wanted to share part of my vision for what our pathforward may look like. Before I dive into that, I briefly wanted to touch onkind of what is MYC, what is its normal structure and what is its function. Andhere on the left, A shows a bio computational method for taking a look at howdisordered a protein structure is. And we see a very squiggly line with thedisorder score on the y axis, which indicates that in various domains, MYC isvery disordered, meaning that their 3d structure is really unstable. The reasonthat that's important is that it has historically made it a really difficulttarget to drug.

I'll share the subsequent sides, ways in which we're getting around that butthat is just to say that it's an intrinsically disordered protein. MYC is partof a super family with three different isoforms or paralogues. c-MYC is the onethat's involved in osteosarcoma. We see some shared domains and MYC, you mayhave heard of come up in neuroblastoma and our l-MYC is involved in lungcancer. C shows an x-ray structure of MYC with its binding partner Max bindingto DNA. In the normal physiologic healthy state, MYC is very tightly regulatedand combined to a broad but selective set of high-affinity target genes,allowing for appropriate transcription and repression of genes that allow cellsto proliferate and die when they're under damage or stress and then overall tocontinue to thrive. But in the deregulated straight state, MYC can kind of gowild, it can willy nilly bind to both the high-affinity target genes and low-affinitytarget genes, resulting in uncontrolled cell division proliferation and tumorgenesis.

This is a classic diagram of the hallmarks of cancer and MYC is really at thecenter of all of them. It is a very smart gene that as we discussed is involvedin sustaining proliferative signals, innovates growth suppression, avoidsimmune destruction, enables replicative immortality, and promotes inflammation,invasion, and metastasis. The development of additional blood vessels orangiogenesis promotes genomic instability and mutations and then escapesprogram cell death or apoptosis. Then it's also involved in deregulatingcellular energetics and metabolism. So, it's involved in many, many cancerpathways. Thus, there are many opportunities to think about, therapeuticallyexploiting this gene.

So, this is my version, my own schematic of a multi omoMYC approach, and whereI think we can and will go with this first. And I think we can and willultimately use MYC as a biomarker for risk stratification in counseling patientsand caring for patients, and then the design of our future clinical trials.We're not that yet there. As I discussed, we don't yet have a copy numberthreshold and this will really necessitate broad data sharing and, and, andvalidation in a prospective manner. I think MYC is not only a driver of anumber of different cancers but really can be used as a driver of biologicalinsights to understand those cancers and develop better approaches to treatingthem. Then, despite the fact that it's been called undruggable, I think thereare many efforts that suggest that it ultimately can be attractable therapeutictarget.

So, for the first aim in terms of incorporating it as a biomarker and reallybringing osteosarcoma into the precision oncology space, this is an area thatI'm passionate about, and I think will really require continued broad datasharing aggregation and harmonization of data. I think, in our games studywhere we had 12 different institutions, and multiple different cohorts of dataover decades, we only had 178 patients, and samples. So, I think we need tocontinue to be vigilant about our data sharing, and really link that omics datato the clinical data to say how do these biological insights really impactpatients and what can we do to bring that into their care? There are a numberof really wonderful initiatives including ACR Zahm, Project Genie, and theCCDI, as well as the targeted initiative that really gets at this, theimportance of linking those omic data with the clinical and then helping us tobring it to bear and patients. I share this slide as sort of a teaser ofbiological insights that beg more questions that will hopefully get someadditional insights. Again, I think we need to study the genomic evolution andunderstand MYC's role in from the primary tumor to the recurrent to themetastatic tumor. I'll skip over that.

These are data from Alejandro Sweet-Cordero's lab of which I'm now a proudmember that I wanted to share, which starts to get at the question of is MYCcopy number everything or, and how does it relate to MYC expression? A justshows from an RNA seek standpoint or expression standpoint, that it isextremely heterogeneous over a number of osteosarcoma cell lines and PDX modelsthat we see, each of these dots are a different cell line. There's a high andlow expression ranging the full range. When we look at the RNA seek landscape,in terms of MYC expression, we again see extreme heterogeneity, and then whenwe break that down and C, to look at MYC high expression versus it's shown inorange and MYC low expression. So, in blue, we see differentially expressedgenes in each of those subgroups, which I think is interesting in terms ofunderstanding potentially distinct groups of biologically, groups ofosteosarcoma sarcomas, that clinically and biologically behave differently andmaybe would respond differently to different treatments. Then D, again, is copynumber changes at the genomic level the same as expression. I think this is aclear answer that no, and here on the x-axis, we have copy numbers less than 2,2 to 4, and then 4. We see a really broad range of expression at the RNA levelwithin each of those groups, and they're overlapping with one another. So,there must be some other mechanisms to regulate the ways in which MYC isactually expressed. It probably behooves us to take into account again, notjust the genome level changes, but how does it relate to transcriptome-relatedchanges and expressions and changes in the expression profile of MYC.

There's been some additional really nice work done by a lab memory. You seeFuentes, who is a wonderful member of the lab, who's sits by my bay, who hasdone some work looking at epigenetically defined clusters based on MYCexpression using a methodology called attack-seek. And I won't go into the fulldetails of the approach here, but what we can see is a red cluster on a bluecluster and the red cluster are those of osteosarcoma cell lines and patientdrives genome graph models that have high MYC expression. They cluster awayfrom those commercial cell lines and those that do not have high MYCexpression. When they then did a transcription factor enrichment analysis tosay, well, how does this actually affect what's downstream of MYC, and lo andbehold, we found different transcription factors that were enriched in each ofthese groups. To the right here in blue, we have the high MYC cluster versusthe low MYC cluster. I think this is really exciting given the really profoundheterogeneity of osteosarcoma, to be able to distill this complexity into twoseemingly biologically and clinically relevant groups, we can then start tothink about integrating these kinds of groups into our preclinical, andultimately, our clinical trials.

Can we drug the undruggable? Well, Nelson Mandela and a lot of scientists sayit always seems impossible until it's done. Indeed, we have many, many effortsagain, MYC is a hotspot for nearly all tumors. So, there are many, many effortsacross cancer research for direct and indirect small-molecule inhibition ofMYC, and here's just an overview of some of those strategies, and I'llhighlight a few of them. As we've shared, MYC needs, its binding partner Max todo its function and so interrupting the binding of MYC to Max is one attractivestrategy. OmoMYC is the most commonly studied in the preclinical setting, butthere's a whole host of others and efforts to bring those into clinical trials.

Targeting Max, Max itself is another strategy targeting Aurora kinase A hasshown, preclinical and some clinical efficacy. And there's a compound calledAlister tube, which is in trial or a kinase a inhibits the degradation of MYCs.So, if you can inhibit our kinase perhaps we can promote the degradation ofMYC. And we're seeing that as a potentially efficacious signal in other tumortypes as well. And then bromodomain inhibitors or Brd for as a subclass, butit's also gained a lot of interest in the preclinical setting with efforts tobring them into trial. If you Google MYC drug, I think you get over 8 millionhits. So, there are lots and lots of people working on this. Highlighting acouple of other potential strategies and sort of a broader view on precisiononcology in the osteosarcoma world. These are this is a nice schematic that wasgenerated by Liane sales and a landmark paper that she and others in thespeaker Darrell lab-produced back in 2019, which show that there arebiologically distinct clusters that can be formed based on specific genomicalterations and that these are differentially responsive to different therapiesin patient-derived xenograft cell lines and in mouse models, and highlightingMC here, we see a differential sensitivity of MC amplified PDX is to two CDKinhibitors specifically about to be related to CDK nine inhibition.

I'll briefly go over some of her really nice in vitro and in vivo work. Theseare just two cell lines that show MYC amplification here. This long bar is MYCin the two different cell lines that are derived from patients' tumors,confirming that MYC is actually expressed as a western blot showing these darkbands with expression. These are the control ones where we knew that MYC wasn'tamplified and indeed we don't see a band there. This, in this cell line, whenit is treated with vehicle or the control we see that it grows out of control,but when it's treated with this Pan-CDK inhibitor, we see a stopping of thegrowth and sort of stagnation of tumor volume here in green. Another way ofseeing that is that we see a significantly smaller tumor volume in the treated grouprelative to the control. In some cases, we see again, a lower rate of growth ora lower change percent change in volume, and in some cases, we even seeshrinkage. So, some exciting data that of course would need to be validated inhumans.

Similarly, on osteosarcoma in the second cell line with MYC amplification, wesee the treated group has significantly smaller tumor volume than those thatare treated with the vehicle. she confirmed using a western blot that this isactually mediated through repressing the expression of MEK, we're with thevehicle-treated group we see then the bands show up, and then when it's treatedwith the drug, the CDK inhibitor, we no longer see make an expression. Here'sas I'm showing that, that treatment with the CDK inhibitor increases apoptosisand decreases proliferation. MYC is thought to be involved in the DNA damageresponse through directly inducing DNA damage with a reactive oxygen speciesand when it's functioning normally actually acts to help upregulate p-53 and allowfor apoptosis, but when it's dysregulated, apoptosis is not allowed to is notable to occur.

We see a specific signature within osteosarcoma, tumors, and cell lines we callsignature 3, that's quite similar to BRCA deficient tumors that are common inbreast and ovarian cancers. That data in combination with some in vitro datashowing some response to DNA damage, targeting agents has formed the basis of anew clinical trial. It's a phase 2 study showing olaparib, which is a PARPinhibitor in combination with Serral assertive, which is an ATR inhibitor. Bothof those are involved in the DNA damage response pathway for patients withrecurrent osteosarcoma. It's possible and there seems to be a signal at leastin the breast and ovarian cancer world in which cancers where there isincreased MYC expression, and they seem to be differentially responsive tothese DNA damage response targeted agents.

Then lastly, but certainly not least, or comprehensively, MYC has also beenshown to be a global regulator of the tumor microenvironment or immuneresponse, including regulation of PDL-1 in expression, checkpoint inhibitorsare a hot topic in a lot of other cancers. And it is possible that throughunderstanding the mechanisms of MYC, what regulates MYC, and what MYCregulates, we may be able to use amino therapeutics to differentially targetmake overexpressing tumors. So, my overall summary and that, again, is just areal bird's eye view, with lots and lots of I could have had an infinite numberof boxes within each of these. There are probably many other big boxes, but Ithink we can and will bring MYC into clinical trials and into standard patientcare as a biomarker for risk stratification through continued broad datasharing and continued efforts for banking, prospective validation as abiomarker, and linking clinical and genomic data. We'll continue to use MYC asa driver of biological insights with a specific focus on understanding its rolein metastasis, DNA damage response, the tumor microenvironment, what regulatesMYC, and what MYC regulates and then as a potential therapeutic target, throughthat busy slide that I showed earlier, and many, many more efforts.

So, with that, I'd like to say a sincere thank you to all of you for listeningand for all you do to make it better for kids with osteosarcoma in theirfamilies, to Katie and Alejandro, who have both been wonderful, wonderfulmentors and to both laboratories, and to the many, many patients and theirfamilies who relentlessly inspire me each and every day. So, with that, I'lltake any questions or comments.

Christina: Thank you so much, Dr. Marinoff, that was wow so much like so thatwas really interesting, really. I am hopeful too because MYC is a tough one,but I love that you were kind of able to help identify possible avenues forattacking MYC. So, that is great. Everyone, feel free to put any questions inthe q&a. If you want to throw some of those out. One of the drugs that youmentioned are inhibitors was the CDK inhibitor. And I was wondering, are thereany drugs that are currently in trial for that or any that are FDA approved?

Dr. Marinoff:
That is a great question. I am not certain whether it's approved in othersettings, as of yet. I don't think that there are any current clinical trialsin patients for osteosarcoma, but I think there's enough promising preclinicaldata that we will get there if we're not there yet.

Christina: Mm-hmm. Then just so I understand, and also, I'm just kind ofthinking, you know, as a parent thinking about, okay, if I see MYC show up onone of our reports, and what that means. So, you were talking about thedifference between just expression versus copy number changes. And so I thinkthere's like a, maybe a couple of ways that people might get to, to understand whetheror not there's kind of MYC involved in their tumor, right. One might be the IHCstaining. And so that just kind of shows expression right now. And then there'slike actually seeing it in, like in a genomic sequencing report. And there evenI guess you would actually need to have the copy number change data to kind ofunderstand, I guess, where you are in that risk threshold? Is that correct? So,it's not just like, do you have MYC amplification? There's kind of that furtherstep of like, well, to what degree?

Dr. Marinoff: Absolutely. I think you are really honing in on it, it is clearas mud, what is MYC amplification? Is all MYC amplification the same? How do wedefine it? In both the research space and the clinical space, I think is reallychallenging and has really important implications for patients. And the waythat we approach their treatment and the way that we approach counseling, I canonly imagine how anxiety-provoking it would be to hear that this is a MYCamplified tumor without putting it in context, in terms of is this, is thisbased on expression? Is this really high MYC amplification? Are we seeing lowMYC amplification signals? So, I think it really really behooves us to figureout what MYC amplification means what level of MYC amplification is probablyokay, and what level of MYC amplification should say let's call in all thetroops and think about the best possible strategies for, for treating thatpatient. I'm hopeful that we will identify at least a relative copy numberthreshold or expression level and that we will not wait until those patientsrelapse or do poorly. But rather say up front, we know that this is a poorprognostic marker. So, we just need to be really aggressive with ourtherapeutic strategies. I think we will get there But I think right now, we arenot quite sure how to interpret it, the increased MYC expression, or exactlywhat to do with it. But I think knowledge is power and we're gaining thatknowledge. And I think there seems to be a consensus among the field, that MYCis of interest, and we need to figure out what it means and what to do with it.

Christina: Yeah, and actually, so right to that point, we actually have aquestion from someone asking that based on this research and analysis, do youfeel like for osteo-genomic sequencing should happen at primary tumor biopsy?And should that be the standard of care, knowing that it could be thisimportant prognostic biomarker?

Dr. Marinoff: Yes, you're asking something that I think that all patientsshould have at least an analysis of whether MYC is present. I think, ifpossible, all patients should have some degree of sequencing of both theirtumor for somatic alterations as well as germline sequencing, since we knowthat a lot of osteosarcomas is related to or a fair proportion is related togermline alterations. I think in this rare tumor that right where we need everybit of knowledge that we can get, it will help us to better understand thetumor and treat patients going forward. So, I think both for the individualclinical care, it's useful to have that that sequencing data, I hope it onlybecomes more useful such that we're actually able to say like, we're going totreat you with regimen X based on your sequencing. I think we will get therebut I think we will only get there if we get more genomic data that's linked tothe patient if that makes sense.

Christina: Yeah, I mean, not only just for it's useful for research like yoursfor us to collect it and to track it longitudinally, even for that, then alsocertainly for like the individual patient case and care. I was just, you know,the whole thing about necrosis because certainly, that does seem like ameaningful marker to understand the response to chemo. But as pointed out, itdoesn't necessarily correlate to cognosis, and it seems like there are theseother biomarkers that might do a better job of that. So, it'd be interesting tosee, hopefully, that kind of happens sooner versus later kind of moving in thatdirection. I was gonna say we had a couple of other questions come in. So, onewas striking that the expression doesn't necessarily track the copy numbervariation and what is known about the mechanisms of MC dysregulation inosteosarcoma because expression doesn't necessarily track with the copy numberchanges.

Dr. Marinoff: I think this is a field of active research that is an area ofreal interest for the Sweet-Cordero lab and for others in the field, I think wedon't know the exact mechanisms of MYC regulation, certainly some of it seemsto be happening at an epigenetic level, there was a recent paper that just cameout earlier this month showing prognosis defined by different methylationgroup. So, I think the methylome is another area of study to integrate with howit relates to gene expression at the RNA level and at the DNA level. So, Idon't know yet, but I am, I share your eagerness to explore what thosemechanisms are, and then ultimately leverage them to try and target MYC.

Christina: And then a couple of questions coming in about treatment-related.So, on that question about the CDK4, someone asked, it looks like palbociclibis an approved CDK for inhibitors and other cancers. Would that be potentiallyrelevant for MYC amplify patients?

Dr. Marinoff: I think we're still, we're still learning for which subset ofthose CDK4/6 and there are some other related compounds that are in thepreclinical pipeline, as well as in the clinical trial landscape for other tumors. I think that there is going to be a subset of osteosarcoma tumors thatare responsive to pablo and related drugs. I'm not sure yet, if it's the MYC amplified subset of it's those with CDK4/6 alterations, RB1 status is important there. So, I think we really need to take into account not just sort of one alteration in a vacuum, but kind of their entire co print or each individual's genomic landscape to really MYC informed decisions about which drug therapymatches well, but we in the preclinical setting are actually looking at CDK4/6and 246 inhibitors right now, in the MYC amplified and non amplified setting.In the present and in the coming weeks and months, I hope to have an answer.

Christina: Right. Then another question, with very high MYC should chemotherapynot be given and some other treatment tried, kind of talking about the risk stratificationand implications for treatment?

Dr. Marinoff: I think that's a really great question. I think we don't haveenough data, unfortunately right now to suggest a better regimen, then map evenfor the MYC amplified subset, and there is a subset of MYC amplifiedosteosarcomas that actually do surprisingly well. So, we don't know the wholestory. I think if we had another regimen that we knew was going to work, Iwould say absolutely. Let's stop doing the same old thing and do somethingdifferent. My hope is that we will design our next phase of clinical trials tosay patients with MYC amplification should be treated with something different,be it the map normal backbone plus something else, or tweaking somethingdifferent to get at their specific biology but I think right now to do that offtrial without evidence for a better regimen is tricky. But again, I'm hopingthat we'll be able to offer at first different experimental therapy andhopefully, therapy that will become integrated into the standard of care downthe line for patients with MYC amplified osteosarcoma.

Christina: Just circling back to that earlier question about, like how commonMYC is and other cancers. I think you said it was about 20%. I know osteo iskind of it's very much its own beast and it sounds so heterogeneous, but justbecause it is in present, so many other cancers in larger numbers, I'm justwondering what we can learn from kind of those studies in that world, giventhat there's so many cases that we can then kind of take that learning andapply it to us to just because just the numbers are greater in terms of justeverything, samples, patients, data, and how that can kind of help our learningfor osteo?

Dr. Marinoff: I couldn't agree more, I think we should be really capitalizingon all of the data that we have and thinking about cancers as in some ways asmolecularly defined subgroups. So, I think every bit of information that welearn about MYC's role and other cancers, we should be then taking back to ourbenches in the osteosarcoma world and into our clinics, and seeing how whetherthat upholds in the osteosarcoma setting, but none of that ought to bediscounted. And I think that drugs targeting MYC are much more likely to bestudied first in MYC amplified, adult, and other more prevalent cancers andthat we should, we should learn from those and we should, we should be quick toact on successes there and think about in cases where it didn't work, why andunderstanding the role and other tumors, probably the ways in which itregulates the tumor microenvironment and one cancer, there's at least I'm overoverlap in the osteosarcoma world. So, I think we should not limit ourselves toour numbers and really leverage everything we have.

Christina: Yeah, thank you for all your work on this and for a reallyinteresting presentation, and most of all, for your commitment to pediatriconcology patients everywhere. More information on this and all osteophytes canbe found on our YouTube channel, on our website,, or on yourfavorite podcast place. And if you registered for this OsteoBites, you'll getan email with all the information and any other links mentioned today. Nextweek, we'll be talking with Dr. Vishu Avutu from MSK, Memorial Sloan Kettering,about an early phase clinical trial for patients with metastatic osteosarcoma,targeting the DNA damage repair pathway. It's a phase 1-2 study of CNC3 plusgemcitabine in relapsed or refractory osteosarcoma.

Thank you again to Dr. Marinoff, and to our sponsor, BTG SpecialtyPharmaceuticals. Thank you so much, Dr. Marinoff, this was really great info,really appreciate it. By the way, another plug for Factor, I think Dr. Marinoffis going to be there. So, more questions about MYC, she's going to be there,you can ask her more questions. Thank you all for joining us today onOsteoBites, and we hope to see you next week when we chat with Dr. Avutu.Thanks so much.


Other blog posts